In plain words

Most of the large studies that combined many trials found that omega-3 made little or no difference to heart attacks, strokes and other cardiovascular events. One combined study linked it with slightly fewer major events, and more so in trials of one form, EPA, given on its own.

The short answer

Five pooled analyses of randomized trials address this question; they may share many of the same trials, so they are not five separate measurements. Four found little or no effect, or no significant association, on their main cardiovascular measure: Abdelhamid and colleagues (2020; 140,482 people), Aung and colleagues (2018; 77,917 people), Rizos and colleagues (2012; 68,680 people) and Hooper and colleagues (2006; 36,913 people). Khan and colleagues (2021; 135,019 people) found omega-3 was associated with fewer major cardiovascular events (relative risk 0.95). Certainty is moderate, because the trial sets are large but the results split by omega-3 type.

Where the evidence sits

Fewer events

◪ Pooled analysis, Khan, 2021 · may overlap studies counted here · EPA plus DHA showed no drop

Little or no difference

◪ Pooled analysis, Abdelhamid, 2020 · may overlap studies counted here · coronary events may fall slightly
◪ Pooled analysis, Aung et al., 2018 · may overlap studies counted here
◪ Pooled analysis, Rizos et al., 2012 · may overlap studies counted here
◪ Pooled analysis, Hooper et al., 2006 · may overlap studies counted here
□ Pooled analysis reported by Linus Pauling Institute · may include trials listed here · no new data

■ measured by the source's own study · ◪ pooled analysis, may overlap studies counted here · □ reported, reviewed or argued, no new data

Evidence weight8181 / 100?

Who says what

SourceFindingDesignPeopleDose and lengthFunding and independence
Abdelhamid et al., 2020Little or no effect; coronary events may fall slightlyCochrane pooled analysis of trials140,4820.5 to over 5 g a day, 12 to 88 monthsNot in ledger
Aung et al., 2018No significant association with major eventsPooled analysis of trials77,917226 to 1,800 mg EPA a day, mean 4.4 yearsNot in ledger
Rizos et al., 2012Not associated with fewer cardiac deaths or strokesPooled analysis of trials68,680Not in ledgerNot in ledger
Hooper et al., 2006No strong evidence of fewer eventsPooled analysis of trials36,913Not in ledgerNot in ledger
Khan et al., 2021Associated with fewer major events, more with EPA alonePooled analysis of trialsUp to 143,5140.4 to 5.5 g a day, median 2.0 yearsNot in ledger
Linus Pauling InstituteDescribes a 2018 Cochrane review suggesting no effectUniversity reference page describing a pooled analysisNot in ledgerTrials of at least 12 monthsNot in ledger
Abdelhamid et al., 2020

Little or no effect; coronary events may fall slightly

Design
Cochrane pooled analysis of trials
People
140,482
Dose and length
0.5 to over 5 g a day, 12 to 88 months
Funding
Not in ledger
Aung et al., 2018

No significant association with major events

Design
Pooled analysis of trials
People
77,917
Dose and length
226 to 1,800 mg EPA a day, mean 4.4 years
Funding
Not in ledger
Rizos et al., 2012

Not associated with fewer cardiac deaths or strokes

Design
Pooled analysis of trials
People
68,680
Dose and length
Not in ledger
Funding
Not in ledger
Hooper et al., 2006

No strong evidence of fewer events

Design
Pooled analysis of trials
People
36,913
Dose and length
Not in ledger
Funding
Not in ledger
Khan et al., 2021

Associated with fewer major events, more with EPA alone

Design
Pooled analysis of trials
People
Up to 143,514
Dose and length
0.4 to 5.5 g a day, median 2.0 years
Funding
Not in ledger
Linus Pauling Institute

Describes a 2018 Cochrane review suggesting no effect

Design
University reference page describing a pooled analysis
People
Not in ledger
Dose and length
Trials of at least 12 months
Funding
Not in ledger

Do the largest pooled analyses find fewer heart events?

Abdelhamid and colleagues (2020), in a Cochrane review, pooled trials in 140,482 people taking 0.5 g to more than 5 g of long-chain omega-3 a day for 12 to 88 months. On all cardiovascular events together they reported:

cardiovascular events (RR 0.96, 95% CI 0.92 to 1.01; 140,482 participants; 17,619 people experienced events in 43 RCTs; high-certainty evidence)

The same review found that more long-chain omega-3 may slightly reduce coronary heart disease events (relative risk 0.91 in 134,116 people; the ledger has no dose or duration for this outcome).

Does EPA on its own give a different result?

Khan and colleagues (2021) pooled trials in up to 143,514 people at 0.4 to 5.5 g a day for a median of 2.0 years. For major adverse cardiovascular events (MACE) across all of them, they wrote:

Omega-3 FA was associated with reducing MACE (-1.0 incident cases per 1000 person-years; RR, 0.95 [0.92–0.98]; p = 0.002; moderate certainty; Appendix p 16)

In Khan's data, EPA alone at 1.8 to 4.0 g a day was associated with a relative risk of 0.78 for major events, while EPA plus DHA at 0.4 to 5.5 g a day showed no drop (relative risk 0.99); the ledger has no n or duration for either group.

Did dose, length or blinding change the result?

The Cochrane review by Abdelhamid and colleagues tested dose and length:

Overall, effects did not differ by trial duration or LCn3 dose in pre-planned subgrouping or meta-regression.

Aung and colleagues (2018) split coronary results by whether people knew what they were taking; the ledger has no n, dose or duration for this split:

open-label trials: RR, 0.85; 99% CI, 0.72-0.99; P = .01; blinded trials: RR, 0.99; 99% CI, 0.91-1.07; P = .69; heterogeneity P = 0.03

Why the sources disagree

The pooled analyses mostly agree on the overall result and differ in how they slice the trials. Khan and colleagues (2021) separated EPA alone from EPA plus DHA, and the drop sat in the EPA trials. Rizos and colleagues (2012) saw an association with cardiac death (relative risk 0.91 in 68,680 patients; the ledger has no dose or duration) that missed their threshold adjusted for testing many outcomes. All five may pool many of the same trials.

What would settle it

A blinded randomized trial with three arms (EPA alone, EPA plus DHA, and a fully described placebo oil) in adults with and without prior heart disease, at fixed doses for at least four years, with major cardiovascular events as the pre-registered main outcome, repeated by a second independent group.

How this page was made

The sources are five pooled analyses of randomized trials, found on PubMed and Europe PMC, and a reference page from the Linus Pauling Institute. Every claim was copied word for word from its source into the Longevity Circle ledger and checked against the original text, and a person approved each one. Certainty follows the GRADE levels used in medical reviews. Drafting was assisted by AI.

Changelog: 4 Oct 2026, first version, 6 sources.